JP7442070B2 - Cleaning route determination device and cleaning route determination method - Google Patents

Description

The present disclosure relates to a cleaning route determining device and a cleaning route determining method.

In large commercial complexes such as shopping malls, the use of floor cleaning technology by robots equipped with self-driving technology is attracting attention in order to compensate for future labor shortages. Furthermore, in recent years, due to rapid advances in technologies such as AI (Artificial Intelligence) and robotics, automation of cleaning work using cleaning devices is progressing in shopping centers with vast site areas. In particular, with the development of AI and sensing technology, automation of cleaning work using cleaning devices has begun to be put into practical use even in curved building structures often found in commercial facilities and complex building structures with many obstacles. This cleaning device is equipped with a camera and a sensor, and if it detects an obstacle while driving automatically, it can autonomously collect dust inside a commercial facility while avoiding the obstacle. For example, Patent Document 1 discloses a control technology for an autonomous cleaning device. According to Patent Document 1, the cleaning device includes a distance sensor, and a control method for changing the moving speed of the cleaning device according to the distance to an object is disclosed. With this technology, even if the distance between the object and the cleaning device becomes less than a threshold value, the cleaning device can continue its cleaning activity without stopping.

JP2018-86218A

However, cleaning equipment is required to complete cleaning work over a large area, such as several thousand square meters, across multiple floors, within a limited time of several hours. In the conventional technology, even if the cleaning device is able to efficiently perform cleaning work even when an obstacle is present, it is difficult for the cleaning device to clean a large floor area in a commercial facility or the like within a predetermined time. For example, commercial facilities such as shopping malls where cleaning equipment is installed have a large site area of tens of thousands of square meters. It is necessary to efficiently clean a very large space. To achieve this, it is difficult for the cleaning equipment to visit all areas to be cleaned, so it is necessary to schedule the cleaning equipment to visit as many high-priority areas as possible within a given period of time. .

The present disclosure provides a technique for efficiently cleaning the interior of a facility with a very large floor area within a predetermined time.

A cleaning route determining device according to an aspect of the present disclosure includes a calculation condition input unit that receives calculation conditions including information inside the facility, and analyzes the behavior of airflow and the behavior of particulates inside the facility based on the calculation conditions. an analysis unit, and creates a dust accumulation map showing one or more dust accumulation locations where dust accumulates and the amount of dust at the one or more dust accumulation locations within the facility, based on the results of the analysis. a map creation unit; a classification unit that selects one or more clusters from one or more clusters inside the facility based on predetermined conditions; and a first route determined from the second plurality of routes. a route calculation unit; and an attribute cluster calculation unit that clusters a plurality of areas included inside the facility into one or more clusters by machine learning, each of the second plurality of routes or a route through which the cleaning unit passes through at least one dust accumulation location included in the selected one or more clusters among the plurality of dust accumulation locations , and the selection by the classification unit is Based on the characteristics of each of the one or more clusters, the number of entrances and exits of the room existing in the area of the cluster to be selected is one, and the width of the passageway of the entrance and exit of the room is a second threshold value. In the following cases, the classification unit does not select the selection target cluster.

Note that this general or specific aspect may be realized by a system, a method, an integrated circuit, a computer program, or a computer-readable recording medium, and may be realized by a device, a system, a method, an integrated circuit, a computer program, and a computer-readable recording medium. It may be realized using any combination of recording media. The computer-readable recording medium includes, for example, a non-volatile recording medium such as a CD-ROM (Compact Disc-Read Only Memory).

In this disclosure, the terms "fine particles" and "dust" are used without particular distinction.

According to the present disclosure, it is possible to determine a route for efficiently cleaning the interior of a facility with a very large floor area within a predetermined time. Further advantages and advantages of one aspect of the disclosure will become apparent from the specification and drawings. Such advantages and/or effects may be provided by each of the several embodiments and features described in the specification and drawings, but not necessarily all are provided in order to obtain one or more of the same features. There isn't.

FIG. 1 is a diagram showing an example of a system configuration in this embodiment. Figure 2 is a diagram showing an example of a bird's-eye view of the internal passageways, room positions, and shapes of the facility. Figure 3 is a diagram showing an example of setting facility entrances and exits, ventilation opening positions, air conditioning equipment positions, and stair positions in drawing information. Figure 4 is a diagram showing an example of a dust accumulation map when a particulate simulation is performed under certain conditions in an example in which facility entrances and exits, ventilation opening positions, air conditioner positions, and stair positions are set in the drawing information. Figure 5 is a diagram showing an example of clustering. FIG. 6 is a diagram illustrating an example of areas to be cleaned by a robot and areas to be cleaned by a person, reflecting the results of attribute determination by the attribute calculation unit 22. Figure 7 shows the results of clustering dust accumulation locations in an example of a dust accumulation map when a particulate simulation is performed under certain conditions in an example in which facility entrances and exits, ventilation opening locations, air conditioner locations, and staircase locations are set in the drawing information. Diagram showing an example of Figure 8 shows the results of clustering dust accumulation locations in an example of a dust accumulation map when a particulate simulation is performed under certain conditions in an example in which facility entrances and exits, ventilation opening locations, air conditioner locations, and staircase locations are set in the drawing information. Diagram showing an example of a cleaning route determined by an optimization algorithm based on an example FIG. 9 is a diagram showing an example of how to display the optimal cleaning route, etc. Figure 10 is a flowchart of the cleaning route calculation system. Figure 11 is a flowchart of the cleaning section. Figure 12 is a flowchart of the terminal.

Hereinafter, embodiments will be specifically described with reference to the drawings.

Note that the embodiments described below are all inclusive or specific examples. The numerical values, shapes, materials, components, arrangement positions and connection forms of the components, steps, order of steps, etc. shown in the following embodiments are examples, and do not limit the present disclosure. Further, among the constituent elements in the following embodiments, constituent elements that are not described in the independent claims indicating the most significant concept will be described as arbitrary constituent elements.

FIG. 1 is a diagram showing an example of a system configuration in this embodiment. The example in this embodiment includes a cleaning route calculation system 1, a cleaning section 2, and a terminal 3.

The cleaning route calculation system 1 includes a dust behavior calculation section 10, an attribute determination section 20, and a cleaning route calculation section 30. The dust behavior calculation section 10 includes a drawing input section 11 , a behavior calculation section 12 , and a dust map creation section 13 . The attribute determination unit 20 includes an attribute cluster calculation unit 21 and an attribute calculation unit 22. The cleaning route calculation section 30 includes a dust cluster calculation section 31 , a cleaning route determination section 32 , and a communication section 33 .

The cleaning section 2 includes a cleaning section 40 and a driving section 50. Further, the cleaning section 40 includes a dust collection section 41 and a dust amount detection section 42. Further, the drive section 50 includes a moving section 51 , a control section 52 , a self-position estimating section 53 , and a communication section 54 .

The terminal 3 includes a dust map display section 60. The dust map display section 60 includes a communication section 61 , a display section 62 , and an input section 63 .

Each part will be explained below.

The cleaning route calculation system 1 performs airflow behavior analysis and particulate behavior analysis in order to determine dust accumulation locations inside the facility. Then, based on the results of airflow behavior analysis and particulate behavior analysis, a dust accumulation map is created that shows where dust has accumulated inside the facility. In addition, calculations are performed to divide the internal areas of the facility into clusters based on their characteristics. The optimal cleaning route is derived from the information on the dust accumulation map and the results of dividing the internal areas of the facility into clusters. Then, the derived optimal cleaning route is transmitted to the cleaning section 2.

The dust behavior calculation unit 10 is composed of a CPU, a memory, and the like. The dust behavior calculation unit 10 is an example of a calculation condition input unit.

The drawing input unit 11 receives input of drawings of facilities and the like. Further, the drawing input unit 11 may include a drawing creation support function that allows the user to create a drawing. The drawing input unit 11 includes a memory, an input terminal, a touch panel, a video output screen, and the like. The drawing input unit 11 may operate using a CPU.

The behavior calculation unit 12 executes air flow simulation and particulate simulation inside the facility for calculating the behavior of dust. The behavior calculation section 12 is an example of an analysis section.

The dust map creation unit 13 creates a dust accumulation map showing the locations where dust has accumulated inside the facility based on the results of the airflow simulation and particulate simulation inside the facility executed by the behavior calculation unit 12. . The dust map creation unit 13 may include an output terminal or the like. The dust map creation unit 13 is an example of a map creation unit.

The attribute determination unit 20 is an example of a classification unit.

The attribute cluster calculation unit 21 performs clustering of passages and/or rooms inside the facility using an algorithm for unsupervised machine learning. Any algorithm may be used as long as it can perform non-hierarchical cluster analysis.

The attribute calculation unit 22 classifies the clusters into clusters that are cleaned by humans and clusters that are cleaned by the cleaning unit 2 based on the calculation results of the cluster calculation unit. The classification is not limited to dividing clusters into clusters where cleaning is performed by a person and clusters where cleaning is performed by the cleaning unit 2. The classification may be performed into three or more categories.

The cleaning route calculation unit 30 is composed of a CPU, memory, and the like. The cleaning route calculation unit 30 is an example of a route calculation unit.

The dust cluster calculation unit 31 performs clustering of dust accumulation locations based on the dust accumulation map created by the dust map creation unit 13. A dust accumulation place is a place where dust is accumulated. The dust cluster calculation unit 31 performs clustering of dust accumulation locations using an algorithm for unsupervised machine learning.

The cleaning route determining unit 32 receives the results of clustering the dust accumulation locations received from the dust cluster calculation unit 31, information indicating the amount of dust accumulated at each dust accumulation location, and the clusters to be cleaned received from the attribute calculation unit 22. Based on the specified information, an optimal cleaning route (optimal route) is derived. The cleaning route determination unit 32 extracts information including dust accumulation locations where the amount of accumulated dust is greater than or equal to a predetermined threshold from the information received from the dust cluster calculation unit 31, and derives a cleaning route that passes through the dust accumulation locations. You can also. By adopting such a configuration, cleaning can be performed more efficiently.

The communication unit 33 transmits the optimal cleaning route derived by the cleaning route determining unit 32 to the cleaning unit 2. Further, the communication unit 33 may transmit the optimal cleaning route (optimal route) derived by the cleaning route determining unit 32 to the terminal 3. The communication unit 33 includes a transmitter, a receiver, and the like. Further, the communication unit 33 may include a cable such as a LAN cable, a NIC (Network Interface Card) for a wireless LAN, and the like. Further, the communication unit 33 may include a wireless communication device or the like.

The cleaning unit 2 is a cleaning device that operates autonomously. The cleaning unit 2 may be an autonomous cleaning robot.

The cleaning unit 2 operates autonomously and cleans the inside of the facility by collecting and removing dust. Further, the cleaning section 2 may perform wiping.

The cleaning section 40 removes dust from the floors, walls, etc. of the facility.

The dust collection unit 41 collects dust accumulated on the floor or wall of a facility by suction or the like. The dust collecting section 41 is configured with a suction device such as a vacuum that sucks dust and the like.

The dust amount detection unit 42 measures the amount of dust collected by the dust collection unit 41. The dust amount detection section 42 includes a weight sensor. The weight sensor may be a load cell or an electromagnetic balance sensor. The dust amount detection section 42 may include various sensors other than the weight sensor.

The drive unit 50 includes a power source such as a motor that moves the cleaning unit 2.

The moving section 51 has a function of moving the cleaning section 2. The moving unit 51 includes a power source such as a motor, wheels, and the like.

The control unit 52 controls the movement of the drive unit 50 as a whole. The control unit 52 includes a CPU, memory, and the like.

The self-position estimating unit 53 grasps the position inside the facility where the cleaning unit 2 is located. The location within the facility may be a relative location or an absolute location. The self-position estimating unit 53 includes a memory, various sensors, and the like. The self-position estimating unit 53 may operate using a CPU.

The communication unit 54 receives the optimal cleaning route derived by the cleaning route determining unit 32 from the cleaning unit 2 . Further, the information collected by the cleaning section 2 may be transmitted to the cleaning route calculation section 1 or the terminal 3. The communication unit 54 includes a transmitter, a receiver, and the like. Further, the communication unit 54 may include a cable such as a Lan cable, a NIC (Network Interface Card) for a wireless LAN, and the like. Further, the communication unit 54 may include a wireless communication device or the like.

The terminal 3 is a portable device with input/output functions. The terminal 3 displays a dust accumulation map and accepts input to the dust accumulation map. Further, the input information etc. are transmitted to the cleaning route calculation system 1 or the cleaning section 2. In addition, information regarding the dust accumulation map or the position of the cleaning unit 2 is received from the cleaning route calculation system 1 or the cleaning unit 2 . The terminal 3 may be a tablet terminal, a smartphone, a PDA, or the like.

The dust map display section 60 displays the dust accumulation map created by the dust map creation section 13 on the terminal 3.

The communication unit 61 receives the dust accumulation map created by the dust map creation unit 13 from the communication unit 33 included in the cleaning route calculation unit 30. The communication unit 61 includes a transmitter, a receiver, and the like. Further, the communication unit 61 may include a cable such as a Lan cable, a NIC (Network Interface Card) for a wireless LAN, and the like. Further, the communication unit 61 may include a wireless communication device or the like.

The display unit 62 displays the dust accumulation map created by the dust map creation unit 13 on a screen or the like. The display section 62 is composed of an organic EL display, a liquid crystal display, or the like. The display unit 62 may include a touch panel or the like.

The input unit 63 receives information such as the location of the dust accumulation area that has been cleaned. The information may be entered manually on the terminal 3, or may be entered via wireless or wired communication from another device. The input unit 63 may include a CPU, memory, communication device, keyboard, etc. The input unit 63 may be configured with a touch panel or the like.

(Cleaning route calculation system)
The functions of each component of the cleaning route calculation system 1 will be explained in more detail.

The cleaning route calculation system 1 performs airflow simulation and particulate simulation inside the facility. The cleaning route calculation system 1 may be provided in a server. The facility is, for example, a commercial facility. Based on the result of the particulate simulation, the dust map creation unit 13 creates a dust accumulation map showing the points where dust is accumulated. Then, each storage location is classified according to rules based on the internal characteristics of the facility, and an attribute is assigned to indicate whether it should be cleaned by a person or by a robot. In addition, the cleaning route calculation unit 30 calculates, from the output dust accumulation map, an optimal route for the cleaning device to visit points where dust has accumulated within a predetermined period of time. The number of points to be visited may be at least one, but may be two or more.

The dust behavior calculation unit 10 creates a calculation model in order to calculate the behavior of dust inside the facility. Therefore, the dust behavior calculation unit 10 uses three-dimensional CAD ( Facility drawing information such as Computer-Aided Design) data is read into the drawing input section 11. If there is no CAD data that can be used by the drawing input unit 11, the drawing input unit 11 launches a drawing input support application on the server, allowing the user to create a drawing.

The above facility drawing information and each of the above drawings may include information specifying the position and shape of a passage inside the facility, and/or information specifying the position and shape of a room inside the facility.

FIG. 2 is a diagram illustrating an example of a bird's-eye view of passages inside the facility, the positions and shapes of rooms. When the drawing input unit 11 receives a drawing input using one of the above methods, the dust behavior calculation unit 10 determines the boundary conditions to be set for the calculation model. 10 accepts inputs such as positional information regarding the positions of facility entrances and exits inside the facility, ventilation holes within the facility, air conditioning equipment, and stairs.

FIG. 3 is a diagram showing an example in which facility entrances and exits, ventilation opening positions, air conditioner positions, and stair positions are set in drawing information. Since dust flows in from outside the facility, the amount of inflow and outflow of particulates is determined by the location of facility entrances and exits that communicate with the outside of the facility and the locations of ventilation openings. At this time, the dust behavior calculation unit 10 may determine the inflow amount of particulates from the number of times the door is opened/closed per day, which is measured by a sensor attached to the door at the entrance/exit of the facility. By doing so, the cleaning route calculation system 1 improves the accuracy of the simulation showing the behavior of dust.

If the sensor attached to the facility entrance/exit door cannot be used, the dust behavior calculation unit 10 calculates the average number of door openings/closings based on the average number of visitors per day, etc. The concentration of incoming and outgoing particulates may be determined. At this time, the dust behavior calculation unit 10 may use a general particulate concentration as a default value as the particulate concentration outside the facility, but the dust behavior calculation unit 10 may use the PM2.5 distribution prediction information for each region published on the web, etc. The particulate concentration outside the facility may be determined based on the The accuracy of the simulation showing the behavior of dust is improved by the dust behavior calculation unit 10 using PM2.5 distribution prediction information for each region published on the Web.

There are two boundary conditions for wind speed at facility entrances and exits, which are used as boundary conditions to determine the concentration of incoming and outgoing particulates: the wind speed of natural convection from outside the facility and the wind speed of airflow induced by people walking. Regarding the wind speed of natural convection from outside the facility, the cleaning route calculation system 1 either measures the wind speed with a sensor before the facility opens, or sets a general wind speed value as a default value. Note that the user may set an arbitrary wind speed value in the dust behavior calculation unit 10. Further, the dust behavior calculation unit 10 determines the wind speed induced by a person's walking based on the walking speed of a typical person. These two factors determine the effective inflow conditions for fine particles at the facility entrance. Further, by constantly measuring the dust behavior calculation unit 10 using a wind speed sensor attached to the entrance/exit of the facility, it is possible to directly measure the effective inflow/outflow conditions of fine particles.

Similarly, the dust behavior calculation unit 10 receives input of conditions related to air inflow and outflow at the ventilation opening. If the dust behavior calculation unit 10 can use data such as environmental measurements of the facility, the dust behavior calculation unit 10 may set the data for the calculation model. If there is no data such as environmental measurements of the facility, the dust behavior calculation unit 10 either uses a wind speed sensor or the like to sense the wind speed in the same way as at the entrance/exit of the facility, or detects the wind speed based on the ventilation volume set for each ventilation opening. Perform a simple simulation and determine the wind speed.

For air conditioners, the wind direction, air volume, etc. are set in advance, so the dust behavior calculation unit 10 sets the set wind direction, air volume, etc. as boundary conditions.

In stairs, particulates flow in and out between the floor that is the target area of calculation and another floor, but the dust behavior calculation unit 10 uses the information from the wind speed sensor as in the case of ventilation openings. Determine the wind speed as a boundary condition using data.

In addition, the initial settings, including information that specifies the position and shape of passages inside the facility, and/or information that specifies the position and shape of rooms inside the facility, and boundary conditions must be set when performing the simulation for the first time. Just go. There is no need to set it again in the second and subsequent simulations. However, the wind speed and the inflow/outflow conditions of particulates may be automatically updated based on sensor information and/or information obtained from the Web. The above method can save the user's effort.

Once information specifying the position and shape of passages inside the facility and/or information specifying the position and shape of rooms inside the facility and boundary conditions are determined, the dust behavior calculation unit 10 first calculates the airflow within the facility. Perform analysis. For the airflow analysis, commercial CFD (Computational Fluid Dynamics) software or a home-made program based on the finite element method and/or finite volume method may be used. When calculating a large area such as a commercial complex, a turbulence model such as the k-ε model may be used to reduce the calculation load. Moreover, since the calculation needs to be completed in a realistic time, the cleaning route calculation system 1 may perform parallelization using a CPU (Central Processor Unit) and/or a GPU (Graphics Processing Unit). At this time, the lattice Boltzmann method may be used as an algorithm suitable for parallelization. By parallelizing using GPU and performing calculations using the lattice Boltzmann method, cleaning route calculation system 1 can speed up calculations by more than 1,000 times compared to sequential calculations using one normal CPU core. can be expected. Therefore, the dust behavior calculation unit 10 can complete the calculation in a relatively short execution time.

Next, when the airflow analysis calculations are completed, the dust behavior calculation unit 10 calculates particle behavior based on the calculated airflow in order to create a dust accumulation map. When analyzing particle behavior, the dust behavior calculation unit 10 uses a model called a one-way model in which the force acting on particles is determined by airflow. By the above method, the dust behavior calculation section 10 can reduce the calculation load. The airflow data used in the one-way model may be calculated by airflow analysis. The dust behavior calculation unit 10 uses forces such as drag and gravity as boundary conditions and the like when analyzing particulates. The dust behavior calculation unit 10 uses forces such as buoyancy as a boundary condition or the like when coupling particle analysis and thermal fluid analysis. Further, regarding the particle size and density, the dust behavior calculation unit 10 determines the particle size distribution by referring to information obtained from the Web, since the typical particle size distribution may differ depending on the region. Alternatively, the particle size and/or density information may be manually set by the user. The dust behavior calculation unit 10 may use values such as a particle size of 2.5 μm and a particle density of 1050 kg/m ³ as default values for the particle size and density.

The dust behavior calculation unit 10 performs this particulate analysis for a predetermined period of time, such as one day, and counts the amount of particulates that have settled on the floor surface of the facility due to gravitational settling during the predetermined period of time. An accumulation map can be created based on the calculation results of the dust behavior calculation section 10.

FIG. 4 is a diagram showing an example of a dust accumulation map when a particulate simulation is performed under certain conditions in an example in which the drawing information includes facility entrances, ventilation opening positions, air conditioner positions, and stair positions. Here, the accumulated amount of particulates is shown in a contour diagram, and this map allows the user of the cleaning route calculation system 1 to grasp in advance the areas in the facility that require cleaning. Further, the dust behavior calculation unit 10 performs this particulate simulation at predetermined time intervals, such as every 12 days, and updates the dust accumulation map. In that case, the dust accumulation map will be updated by using the dust accumulation map from the previous simulation as the initial value and adding the calculated value of the dust accumulation map from the current simulation to the initial value. . The dust accumulation map created here is passed from the dust behavior calculation section 10 to the cleaning route calculation section 30. Note that the dust map creation section 13 may transmit the created dust accumulation map to the terminal 3 through the communication section 33.

Next, in the attribute determination unit 20, the attribute cluster calculation unit 21 performs clustering of passages and/or rooms within the facility. The attribute cluster calculation unit 21 uses one of various algorithms for clustering, such as the k-means method, which is an unsupervised machine learning algorithm, support vector machine, and random forest. The inputs to the algorithm are information indicating the planar shape of a plurality of rooms inside the facility and the positions of the plurality of rooms, and information indicating the planar shape of a plurality of passages inside the facility and the positions of the plurality of passages. The output from may be an index for each of a plurality of rooms and an index for each of a plurality of hallways.

FIG. 5 is a diagram showing an example of clustering. According to FIG. 5, it can be seen that the attribute cluster calculation unit 21 clusters the passages inside the facility according to the narrowness of the passage or the number of turns. In order to make the display easier to understand, rooms are not clustered here, but clustering may also be performed in areas inside the facility other than aisles such as stores and restrooms.

Next, the clustering results are passed to the attribute calculation unit 22, and it is determined whether the cleaning is to be performed by a person or by the cleaning unit 2, depending on the characteristics of the cluster. Robotics technology and sensing technology do not have sufficient accuracy for the cleaning unit 2 to move and detect obstacles. For this reason, there is a possibility that the cleaning unit 2 may drop the products while cleaning the area where the products are displayed. Further, in an area like the cluster 3 in FIG. 5, which is not long but has many turns, the cleaning unit 2 needs to rotate at each turn, which is likely to waste time. On the other hand, large aisles have fewer obstacles, but occupy most of the area within a commercial facility. It is more efficient to clean these areas by the cleaning unit 2.

The attribute calculation unit 22 determines the attributes of each cluster according to the characteristics of each cluster, and assigns attribute information indicating the attributes to each cluster. The attribute calculation unit 22 removes clusters that should not be cleaned by the cleaning unit 2 from all the clusters, and passes information specifying clusters that the cleaning unit 2 should clean to the cleaning route determining unit 32. The information specifying the cluster to be cleaned may be information specifying the robot cleaning area shown in FIG. 6. The information specifying the cluster to be cleaned is a set of information specifying each of cluster 1, cluster 2, cluster 4 to cluster 10 shown in FIG. 5, that is, {set specifying cluster 1, information specifying cluster 2 , information specifying cluster 4 to information specifying cluster 10}. When a drawing that specifies the internal structure of the facility (i.e., the position and shape of passages inside the facility, and/or the position and shape of rooms inside the facility) is described on a two-dimensional plane, information that identifies cluster i may be a set of information indicating coordinates included in cluster i in the two-dimensional plane. This prevents the cleaning route calculating section 30 from falling into a local solution when calculating the optimal cleaning route (optimum route), and makes the cleaning work by the cleaning section 2 effective. The optimal route) can be derived.

FIG. 6 is an example showing an area to be cleaned by a robot and an area to be cleaned by a person, reflecting the result of attribute determination by the attribute calculation unit 22.

In determining the attributes, the attribute calculation unit 22 determines the attributes as follows, taking into account the nature of the commercial facility.

1. Whether inside a store: Inside a store, there is a possibility that the cleaning department 2 may drop the displayed products during cleaning work. For this reason, it is more appropriate for the inside of the store to be cleaned manually. When the cluster is a store, the attribute calculation unit 22 determines that the cluster is an area that should be cleaned by a person. That is, the attribute calculation unit 22 does not select the cluster as a cluster to be cleaned by the cleaning unit 2.

2. The number of curved corners in the area is greater than or equal to the threshold If there are a large number of curved corners inside the cluster, the cleaning unit 2 needs to change direction many times in order to clean that area, and even if the cleaning distance is short, Cleaning by the cleaning section 2 takes time. For this reason, such an area is not suitable for cleaning by the cleaning unit 2, and it is more appropriate to clean it manually. For example, eight may be used as the threshold for the number of turns per region. When the number of curved corners existing in the area of a cluster is equal to or greater than a threshold value (e.g., eight), the attribute calculation unit 22 determines that the cluster is an area that should be cleaned by a person. That is, the attribute calculation unit 22 does not select the cluster as a cluster to be cleaned by the cleaning unit 2.

3. A room where the number of entrances and exits is only one, and the width of the passageway is less than a threshold value.For example, a toilet usually has only one entrance and exitway, and the passageway of the entranceway is narrow. For this reason, when the cleaning unit 2 enters the room, there is a possibility that the user will not be able to leave the room. For this reason, if the area is inside such a closed facility and the width of the entrance/exit passage is less than the threshold value, cleaning by the cleaning unit 2 is not appropriate and it is more appropriate to do it manually. . However, if the entrance is widened due to renovation or the like, or if the passage becomes sufficiently wide due to renovation or the like, cleaning by the cleaning section 2 becomes possible. For example, 100 cm or the like may be used as the threshold for the width of the passage. The attribute calculation unit 22 calculates that if the number of entrances and exits of a room existing in the area of a cluster is one, and the width of the passageway of the entrance and exit of the room is less than or equal to a threshold value (for example, 100 cm), the cluster is Determine the area to be cleaned. That is, the attribute calculation unit 22 does not select the cluster as a cluster to be cleaned by the cleaning unit 2.

4. The width of the aisle is below the threshold There are places within the facility where the width of the aisle is narrow. In such a narrow passage, there is a high possibility that the cleaning unit 2 will need to change direction many times to avoid collision with the wall. For this reason, such an area is not suitable for cleaning by the cleaning unit 2 and is more suitable for cleaning by a person. Although 100 cm or the like may be used as the threshold for the width of the passage, for example, 90 cm or the like may be used. When the minimum value of the width of the passage existing in the area of a cluster is less than or equal to a threshold value (for example, 90 cm), the attribute calculation unit 22 determines that the cluster is an area that should be cleaned by a person. That is, the attribute calculation unit 22 does not select the cluster as a cluster to be cleaned by the cleaning unit 2.

5. Place where there is a step in the area If there is a step in the middle of the route, the movement of the cleaning unit 2 may stop. Such areas are not suitable for cleaning by the cleaning unit 2 and are more suitable for cleaning by humans. If there is a step within the area of the cluster, the attribute calculation unit 22 determines that the cluster is an area that should be cleaned by a person. That is, the attribute calculation unit 22 does not select the cluster as a cluster to be cleaned by the cleaning unit 2.

6. An area where the physical distance from the cleaning unit 2 is greater than a threshold If the amount of accumulated dust in the area is sufficiently large, the area can be designated as the cleaning route even if it is located far from the starting point of the cleaning unit 2. There is sex. In this case, although the amount of dust collected and recovered is optimal, there is a possibility that many other dust accumulation locations will not be cleaned by the cleaning unit 2. Therefore, if the distance from the starting point of the cleaning section 2 is equal to or more than the threshold value, it is more appropriate for the cleaning to be carried out by a person. The threshold value of the physical distance to the cleaning unit 2 may be, for example, 700 m in straight line distance. If the distance between the center of gravity of the cluster area and the cleaning start position is equal to or greater than a threshold value (for example, 700 m in straight line distance), the attribute calculation unit 22 determines that the cluster is an area that should be cleaned by a person. That is, the attribute calculation unit 22 does not select the cluster as a cluster to be cleaned by the cleaning unit 2.

7. The deviation of the particle distribution within the cluster is greater than the threshold Even if the amount of accumulated dust is the same, if dust is concentrated in a specific place within the cluster, even if a person cleans it, the cleaning part 2 can be cleaned in a short period of time. On the other hand, if dust is evenly accumulated in the cluster, the area that needs to be cleaned becomes large, so cleaning takes time whether it is done by a person or by the cleaning unit 2. In such a case, the latter may be cleaned by the cleaning section 2 from the viewpoint of reducing the burden on the cleaning staff. For this reason, if the deviation of the particle distribution within the cluster is equal to or greater than the threshold value, it is more appropriate to clean it manually. The attribute calculation unit 22 may use the standard deviation of the particle distribution within the cluster to determine the bias. When the bias in the distribution of particulates within the region of a cluster is equal to or greater than a threshold value, the attribute calculation unit 22 determines that the cluster is an area that should be cleaned by a person. That is, the attribute calculation unit 22 does not select the cluster as a cluster to be cleaned by the cleaning unit 2.

8. Places where the traffic of people is more than a threshold The cleaning route calculation system 1 may estimate the flow line within the facility from the number of visitors to the facility and/or each store. In this case, the frequency of people coming and going on each floor or each hour is used to estimate the flow line within the facility. In places with a lot of foot traffic, it is difficult for cleaning staff to perform cleaning work during opening hours due to congestion. Therefore, in the above areas, cleaning may be performed by the cleaning unit 2 at night. Therefore, when the frequency of people coming and going in the target area is equal to or higher than the threshold value, it is more appropriate for the cleaning unit 2 to clean the inside of the target area. When the frequency of people coming and going to a cluster area is equal to or higher than a threshold value, the attribute calculation unit 22 determines that the cluster is an area that the cleaning unit 2 should clean. That is, the attribute calculation unit 22 selects the cluster as the cluster that the cleaning unit 2 should clean.

9. Locations that do not fall under any of 1 to 7 In this case, it is more appropriate for the cleaning department 2 to clean the area than by a person.

The rules 1 to 9 above are examples of attribute determination, and other rules may be set as necessary. Calculations related to this attribute determination will be performed as long as the information specifying the position and shape of passages inside the facility and/or the position and shape of rooms inside the facility do not change due to facility renovation or other reasons. You only need to run it once at the beginning.

Figure 7 shows the results of clustering dust accumulation locations in an example of a dust accumulation map when a particulate simulation is performed under certain conditions in an example in which facility entrances and exits, ventilation opening locations, air conditioner locations, and staircase locations are set as drawing information. It is a figure showing an example. The cleaning route calculation unit 30 passes the dust accumulation map calculated by the dust behavior calculation unit 10 to the dust cluster calculation unit 31, and performs clustering of dust accumulation locations. The dust cluster calculation unit 31 uses one of various algorithms for clustering, such as the k-means method, which is an unsupervised machine learning algorithm, support vector machine, and random forest. Because of this clustering, each dust accumulation location belongs to one of the clusters, so a discrete index is assigned to each dust accumulation location.

The result of clustering the dust accumulation locations by the dust cluster calculation unit 31 is {(information identifying the first dust accumulation location, index of the first dust accumulation location), (information identifying the second dust accumulation location, index of the second dust accumulation location), . . . } may be used. When a drawing showing the internal structure of a facility is described on a two-dimensional plane, the information specifying the i-th dust accumulation location may be coordinates indicating the area center of gravity of the cluster i.

The dust cluster calculation unit 31 may generate information by adding information indicating the amount of dust accumulated at each dust accumulation place to the result of clustering the dust accumulation places. In other words, the dust cluster calculation unit 31 calculates {(information specifying the first dust accumulation place, amount of dust accumulation in the first dust accumulation place, index of the first dust accumulation place), (second dust accumulation place), }, the amount of dust accumulated at the second dust accumulation location, the index of the second dust accumulation location), etc. may be generated.

The cleaning route determination unit 32 specifies the clustering results of the dust accumulation locations received from the dust cluster calculation unit 31 and information indicating the amount of dust at each dust accumulation location, and the clusters to be cleaned received from the attribute calculation unit 22. Based on the information provided, cleaning routes are determined. When cleaning inside a facility, the cleaning department 2 needs to complete the cleaning work within the vast facility grounds within a limited time, such as 4 to 5 hours. For this reason, the cleaning unit 2 cannot clean all the dust accumulation locations shown in the dust accumulation map. For this reason, the cleaning route determining unit 32 solves an optimization problem of maximizing the amount of cleaning within a predetermined time constraint. Many optimization algorithms are used for such optimization problems. Most of the optimization algorithms used are discrete optimization methods, and as determined by the dust cluster calculation unit 31, it is necessary to assign discrete indexes to dust accumulation locations in advance. The algorithm used by the cleaning route determining unit 32 for optimization is assumed to be a genetic algorithm, simulated annealing, quantum annealing, or the like.

The cleaning route determination unit 32 determines that each of the “information specifying the first dust accumulation place, information specifying the second dust accumulation place, etc.” received from the dust cluster calculation unit 31 is assigned to the attribute calculation unit 22. Check whether it is included in the "information identifying clusters to be cleaned" received from. If the information that specifies the cluster to be cleaned includes information that specifies the dust accumulation place, then the (information that specifies the dust accumulation place, amount of dust accumulated at the dust accumulation place, index of the dust accumulation place) becomes A set of information {(Dust accumulation location ₁ included in the cluster to be cleaned, amount of accumulated dust in the dust accumulation location ₁ included in the cluster to be cleaned, dust accumulation included in the cluster to be cleaned) (Index of location ₁ ), (Dust accumulation location ₂ included in cluster to be cleaned, amount of accumulated dust in dust accumulation location ₂ included in cluster to be cleaned, index of dust accumulation location ₂ included in cluster to be cleaned) ,...} is determined. This information is used to utilize the optimization algorithm described above.

Note that if there is an algorithm that performs continuous optimization instead of the discrete optimization method as described above, the cleaning route determining unit 32 may perform route optimization using the algorithm. In that case, the cleaning route calculation unit 30 does not need to perform clustering of dust accumulation locations.

When determining a cleaning route, the user inputs into the cleaning route calculation system 1 the starting point of the cleaning section 2, the point where the cleaning ends, and the time required to complete the cleaning. Using the above information, the cleaning route determination unit 32 derives an optimal cleaning route that maximizes the amount of cleaning within a predetermined time by solving a constrained discrete optimization problem using an optimization algorithm. Here, maximizing the amount of cleaning within a predetermined time does not necessarily mean maximizing it strictly. When maximizing, a value near the theoretical maximum value may be adopted. Moreover, maximization may be local maximization.

Figure 8 shows the results of clustering dust accumulation locations in an example of a dust accumulation map when a particulate simulation is performed under certain conditions in an example in which facility entrances and exits, ventilation opening locations, air conditioner locations, and staircase locations are set in the drawing information. It is a figure which shows an example of the cleaning route determined by the optimization algorithm based on an example. Here, the cleaning route calculation unit 30 sets the same location as the starting point and ending point of the cleaning unit 2. Figure 8 shows that under the cleaning time constraints, there is no route that can clean all dust accumulation areas, and areas that deviate from the cleaning route remain as areas that require cleaning by cleaning staff. . This calculated optimal cleaning route is sent to the cleaning section 2 via the communication section 33.

(Cleaning Department)
The cleaning section 2 has a cleaning section 40 and a driving section 50. The cleaning section 40 has a dust collection section 41 that cleans dust accumulated on the floor or wall surface of the facility. The dust collected by the dust collecting section 41 is stored inside the dust collecting section 41. A dust amount detection section 42 inside the dust collection section 41 can measure the amount of collected dust.

The cleaning unit 2 can be moved to a target cleaning location by the drive unit 50. The moving part 51 is for moving the cleaning part 2, and is a general moving means such as wheels. The control unit 52 controls the moving unit 51 to move the cleaning unit 2 to an arbitrary location by controlling the amount of rotation of the wheels and the direction change (rotation in the horizontal direction) of the cleaning unit 2. At this time, the self-position estimating section 53 estimates the current position of the cleaning section 2 by reading the rotation angle of the wheels and the cleaning section 2 by the control section 52 and using the above information as input information. A data assimilation algorithm such as a Kalman filter may be used for self-position estimation. Note that the cleaning unit 2 may include a light source such as a laser, and the position may be estimated from the ToF (Time of Flight) of the laser. The communication unit 54 receives information regarding the optimal cleaning route calculated by the cleaning route calculation system 1 and sends the information to the control unit 52. By combining the received optimal cleaning route and the estimated position of the cleaning section, the cleaning section 2 can perform cleaning work along the optimal cleaning route. Furthermore, the dust amount detection unit 42 can also determine where dust has accumulated in the facility and the amount of accumulated dust. During cleaning or at a predetermined timing after cleaning is completed, the cleaning unit 2 checks the location where the cleaning unit 2 actually performed cleaning work and the amount of dust collected by the cleaning unit 2 detected by the dust amount detection unit 42. The information is sent to the cleaning route calculation system 1. The dust map creation unit 13 reflects information regarding the location where the cleaning work was performed and the amount of dust collected on the dust accumulation map. Specifically, in the generated dust accumulation map, the dust map creation unit 13 displays the area designated as the cleaning route and the area that the cleaning unit 2 actually passes through so that it can be distinguished from other areas. Good too. For example, the dust map creation unit 13 may gray out the area through which the cleaning unit 2 has passed on the dust accumulation map or erase it from the dust accumulation map. The dust map creation unit 13 may generate data on the dust accumulation map in which a method for displaying cleaned areas is different from a method for displaying areas that have not been cleaned yet. The dust map creation unit 13 performs calculations to create the dust accumulation map again based on the information sent from the cleaning unit 2 regarding the location where the cleaning unit 2 actually performed cleaning work and the amount of dust collected. may be performed to update the dust accumulation map. The updated dust accumulation map may be transmitted to the cleaning unit 2 and the terminal 3 through the communication unit 54 and the communication unit 61.

FIG. 9 is a diagram illustrating an example of a method of displaying optimal cleaning routes and the like. As shown in FIG. 9, the cleaning route calculation system 1, for example, leaves the contour map displayed at the initial setting on the place or route cleaned by the cleaning unit 2, and displays the contour map in the area where cleaning was not performed. Generate a dust accumulation map with the display changed from broken lines to solid lines. The updated dust accumulation map is then transmitted to the terminal 3 or the cleaning route calculation system 1 through the communication unit 33. The terminal displays an updated dust accumulation map.

By doing this, the cleaning staff can identify cleaned areas and uncleaned areas on the dust accumulation map, thereby increasing the cleaning efficiency of the cleaning staff.

(terminal)
The terminal 3 is used by a commercial facility manager or a cleaning staff in the facility. The terminal 3 displays the dust accumulation map sent from the cleaning section 2 on the display section 62. For example, the cleaning unit 2 performs cleaning at night, and the next morning, the cleaning staff confirms the dust accumulation map that reflects the cleaning results, confirms areas that require cleaning by the cleaning staff, and begins cleaning activities. In this way, areas that require cleaning by cleaning staff are visualized on the dust accumulation map. The terminal 3 also includes an input section 63 for reflecting the cleaning results of the cleaning staff. The cleaning staff inputs information regarding the area cleaned by the cleaning staff into the cleaning route calculation system 1 from the input unit 63. The dust map creation unit 13 sets the amount of dust remaining in the area actually cleaned by the cleaner to zero and reflects it on the dust accumulation map. As a result, the information regarding the latest dust accumulation location is updated on the dust accumulation map, and the information regarding the latest dust accumulation location is used by the cleaning route calculation system 1 as the initial condition of the particle accumulation state in the particle simulation.

The overall flow of the system will be explained below using a flowchart.

(Cleaning route calculation system)
FIG. 10 is a flowchart of the cleaning route calculation system.

First, when the cleaning route calculation system is powered on, the cleaning route calculation system 1 determines whether initial settings are required (step S100). If the cleaning route calculation system 1 has already been calculated and initial settings of the cleaning route calculation system 1 are not required (No in step S100), the cleaning route calculation system 1 calculates the updated dust that reflects the previous cleaning results. The accumulation map is read (step S101). When the cleaning route calculation system 1 needs to be initialized (Yes in step S100) because there is no readable data such as the previous calculation result, such as an updated dust accumulation map, such as at the time of initial startup (Yes in step S100), the cleaning route calculation system 1 performs initial settings.

Next, the drawing input unit 11 determines whether the drawing input unit 11 holds information specifying the internal structure of the facility, for example, drawing information of the facility (step S102). If there is facility drawing information (Yes in step S102), the cleaning route calculation system 1 reads the drawing (step S104). If there is no drawing of the facility (No in step S102), the drawing input unit 11 starts a drawing input support application on the server, and the user creates a drawing using the drawing input support application. The drawing input unit 11 reads the created drawing (step S103). When creating a drawing, a support creation screen is activated in the cleaning route calculation system 1, and the user creates the drawing.

After reading the drawing, it is checked whether attribute information indicating an attribute has been added to each cluster (step S105). If attribute information is not provided (No in step S105), the attribute cluster calculation unit 21 performs clustering of each passageway and/or each room in the facility based on the drawing information (step S106). ).

The attribute calculation unit 22 determines the cluster attribute for each clustered cluster, and assigns attribute information indicating the attribute to each cluster (step S107). The attribute calculation unit 22 determines which areas should be cleaned by the cleaning unit 2 and which areas should be cleaned by a person for the cluster, based on the characteristics of the cluster.

If there is attribute information (Yes in step S105), the attribute calculation unit 22 reads the attribute information (step S108). Note that step S108 may be performed immediately before step S116, which will be described later.

The behavior calculation unit 12 sets boundary conditions such as the position of the ventilation opening, the entrance and exit of the facility, the position of the stairs, the position of the air conditioning equipment, and the respective wind speeds and pressures (step S109).

After setting the boundary conditions, the behavior calculation unit 12 creates a calculation model from information specifying the position and shape of a passage inside the facility and/or information specifying the position and shape of a room inside the facility and the boundary conditions. Create (step S110).

The behavior calculation unit 12 performs an airflow analysis inside the facility (step S111). After the airflow analysis is completed, the behavior calculation unit 12 couples the results of the airflow analysis with a model for analyzing the behavior of particles, and performs behavior analysis of the particles (step S112). Thereby, the behavior calculation unit 12 can grasp the behavior of particulates inside the facility and the location where particulates accumulate on the floor.

When this particle analysis is completed, the behavior calculation unit 12 analyzes the amount of particles that have settled to the floor by gravity based on the result of the particle analysis. Based on this analysis result, the dust map creation unit 13 creates a dust accumulation map (step S113). The dust cluster calculation unit 31 performs clustering of dust accumulation locations using this dust accumulation map (step S114).

The cleaning route determining unit 32 receives an input of the cleaning time of the cleaning unit 2 from the user (step S115). The cleaning route determination unit 32 receives the cleaning time, the results of clustering the dust accumulation locations received from the dust cluster calculation unit 31 and information indicating the amount of dust at each dust accumulation location, and the information to be cleaned received from the attribute calculation unit 22. An optimal cleaning route (optimum route) is determined based on the information specifying the cluster (step S116). That is, the cleaning route determination unit 32 determines a route that is an optimal solution for the purpose of maximizing the efficiency of cleaning work under the condition that each dust accumulation place is passed within a predetermined time.

Then, the cleaning route calculation system 1 transmits the determined cleaning route (optimal route) to the cleaning unit 2 (step S117).

(Autonomous cleaning device)
FIG. 11 is a flowchart of the cleaning section 2.

The cleaning unit 2 determines whether or not it has received the route sent from the cleaning route calculation system 1 (step S200). If the cleaning unit 2 has not received the route (No in step S200), it continues to determine whether the route has been received. If the cleaning unit 2 receives the route (Yes in step S200), the cleaning unit 2 starts cleaning work based on the received route.

Next, the cleaning unit 2 moves to a point included in the cleaning route (step S201). The cleaning unit 2 then removes dust from the location to which the robot has moved (step S202).

The cleaning unit 2 then determines the amount of dust removed (step S203). Thereafter, the location of the cleaning section 2 is estimated based on the amount of rotation of the wheels included in the moving section 51 and the sensing results by various sensors included in the cleaning section 2 (step S204).

Then, when the cleaning unit 2 finishes the cleaning work, it transmits the cleaning result to the terminal 3 (step S205). Note that since it is assumed here that the cleaning unit 2 is driven at night, the final results are transmitted, but when the cleaning unit 2 is driven during the day, a step of transmitting intermediate results is not included. But that's fine.

(terminal)
FIG. 12 is a flowchart of the terminal 3.

The terminal 3 displays a dust accumulation map that reflects the cleaning results sent from the cleaning section 2. First, the terminal 3 receives the optimal cleaning route (optimum route) transmitted from the cleaning route calculation system 1 (step S300). Next, the terminal 3 displays the area to be cleaned that has not been cleaned by the cleaning unit 2 (step S301). Then, after cleaning the area that has not been cleaned by the cleaning unit 2 and that requires cleaning by the cleaning staff, the cleaning staff registers the cleaning result of the cleaning staff in the terminal 3 (step S302). Then, the terminal 3 transmits the cleaning result of the cleaning staff to the cleaning route calculation system 1 (step S303).

(Effects, etc.)
A cleaning route determining device according to an aspect of the present disclosure includes a calculation condition input unit that receives calculation conditions including information inside the facility, and analyzes the behavior of airflow and the behavior of particulates inside the facility based on the calculation conditions. an analysis unit, and creates a dust accumulation map showing one or more dust accumulation locations where dust accumulates and the amount of dust at the one or more dust accumulation locations within the facility, based on the results of the analysis. a map creation unit; a classification unit that selects one or more clusters from one or more clusters inside the facility based on predetermined conditions; and a first route determined from the second plurality of routes. a route calculation unit, each of the second plurality of routes cleaning at least one dust accumulation place included in the selected one or more clusters among the one or more dust accumulation places. This is the route that the part passes within a predetermined period of time.

Thereby, the cleaning route determining device according to one aspect of the present disclosure determines a first route in which the cleaning unit passes through at least one dust accumulation location included in one or more selected clusters within a predetermined time. . By cleaning along this route, the cleaning section can efficiently clean the inside of the facility within a predetermined time.

In the cleaning route determining device according to one aspect of the present disclosure, the total amount of dust in all dust accumulation locations included in the first route is the largest among a plurality of totals, and each of the plurality of totals is This is the total amount of dust at all dust accumulation locations included in each of the second plurality of routes.

As a result, the cleaning route determining device according to one aspect of the present disclosure can provide the cleaning section with a cleaning route that can increase the amount of dust, and the cleaning section can efficiently clean the inside of the facility by cleaning along this route. You can.

The cleaning route determining device according to an aspect of the present disclosure further includes an attribute cluster calculation unit that clusters a plurality of areas included inside the facility into one or more clusters by machine learning, and the cleaning route determination device according to an aspect of the present disclosure is based on the characteristics of each of the one or more clusters.

As a result, the cleaning route determining device according to one aspect of the present disclosure classifies areas inside the facility according to characteristics, and identifies areas with a high possibility of dust accumulation that are appropriate for cleaning by the cleaning unit. By selecting efficiently, it is possible to efficiently clean areas within the facility where dust is likely to accumulate.

In the cleaning route determining device according to one aspect of the present disclosure, when the cluster to be selected is a store, the classification unit does not select the cluster to be selected.

Thereby, the cleaning route determining device according to one aspect of the present disclosure can prevent the cleaning unit from dropping products in the store, and can perform cleaning with high satisfaction for the user.

In the cleaning route determining device according to one aspect of the present disclosure, the classification unit does not select the selection target cluster when the number of turns existing in the area of the selection target cluster is equal to or greater than a first threshold.

Thereby, the cleaning route determining device according to one aspect of the present disclosure prevents the cleaning efficiency from decreasing by prohibiting the cleaning unit from cleaning an area where the number of turns is greater than a predetermined value.

In the cleaning route determining device according to one aspect of the present disclosure, the number of entrances and exits of the room existing in the area of the cluster to be selected is one, and the width of the passage of the entrance and exit of the room is a second threshold value. In the following cases, the classification unit does not select the selection target cluster.

As a result, the cleaning route determining device according to one aspect of the present disclosure can improve cleaning efficiency by prohibiting the cleaning unit from cleaning a room that has one entrance and exit and the passageway width of the entrance is less than or equal to a predetermined width. prevent it from falling.

In the cleaning route determining device according to one aspect of the present disclosure, when the minimum value of the width of passages existing in the area of the selection target cluster is equal to or less than a third threshold, the classification unit selects the selection target cluster. Not selected.

Thereby, the cleaning route determining device according to one aspect of the present disclosure prevents the cleaning efficiency from decreasing by prohibiting the cleaning unit from cleaning a passage whose width is less than or equal to a predetermined value.

In the cleaning route determining device according to one aspect of the present disclosure, if a step exists within a region of the cluster to be selected, the classification unit does not select the cluster to be selected.

Thereby, the cleaning route determining device according to one aspect of the present disclosure prevents the cleaning unit from cleaning the area where the step exists, thereby preventing the cleaning efficiency from decreasing.

In the cleaning route determining device according to one aspect of the present disclosure, if the distance between the center of gravity of the area of the cluster to be selected and the cleaning start position is equal to or greater than a fourth threshold, the classification unit may Do not select a cluster.

Thereby, the cleaning route determining device according to one aspect of the present disclosure prevents the cleaning efficiency from decreasing by prohibiting the cleaning unit from cleaning an area that is more than a predetermined distance away from the cleaning start position.

In the cleaning route determining device according to one aspect of the present disclosure, the classification unit does not select the selection target cluster when the bias in the distribution of the fine particles in the area of the selection target cluster is equal to or greater than a fifth threshold. .

Thereby, the cleaning route determining device according to one aspect of the present disclosure prevents cleaning efficiency from decreasing by prohibiting cleaning of areas where the bias in the distribution of particulates is equal to or greater than a predetermined value.

In the cleaning route determining device according to one aspect of the present disclosure, the classification unit selects the selection target cluster when the frequency of people coming and going to the area of the selection target cluster is equal to or higher than a sixth threshold.

Accordingly, in the cleaning route determining device according to one aspect of the present disclosure, the cleaning unit can perform cleaning that is highly satisfying for the user by prohibiting cleaning of areas where the frequency of human traffic is less than a predetermined value. It can be done.

A cleaning route determining method according to an aspect of the present disclosure includes a calculation condition input step of accepting calculation conditions including information inside the facility, and analyzing the behavior of airflow and the behavior of particulates inside the facility based on the calculation conditions. creating a dust accumulation map showing one or more dust accumulation locations where dust accumulates and the amount of dust at the one or more dust accumulation locations within the facility, based on the analysis step and the result of the analysis; a map creation step, a classification step of selecting one or more clusters from one or more clusters inside the facility based on predetermined conditions, and determining a first route from a second plurality of routes. a step of calculating a route, each of the second plurality of routes cleaning at least one dust accumulation place included in the selected one or more clusters among the one or more dust accumulation places. This is the route that the part passes within a predetermined period of time.

Accordingly, the cleaning route determining method according to one aspect of the present disclosure determines a first route in which the cleaning unit passes through at least one dust accumulation location included in one or more selected clusters within a predetermined time. . By cleaning along this route, the cleaning section can efficiently clean the inside of the facility within a predetermined time.

Although the cleaning route determining device and the cleaning route determining method have been described above based on the embodiments, the present disclosure is not limited to these embodiments. As long as it does not depart from the spirit of the present disclosure, various modifications that can be thought of by those skilled in the art to this embodiment, and forms constructed by combining components of different embodiments are also within the scope of one or more aspects. may be included within.

All or some of the functional blocks included in the block diagram of the cleaning route calculation system 1, the block diagram of the cleaning unit 2, and the block diagram of the terminal 3 shown in FIG. 1 include semiconductor devices, semiconductor integrated circuits (ICs), Alternatively, it may be implemented by one or more electronic circuits including large scale integration (LSI). The LSI or IC may be integrated into one chip, or may be configured by combining a plurality of chips. For example, functional blocks other than the memory element may be integrated into one chip. Here, they are called LSI or IC, but the name changes depending on the degree of integration, and may be called system LSI, VLSI (very large scale integration), or ULSI (ultra large scale integration). Field Programmable Gate Arrays (FPGAs), which are programmed after the LSI is manufactured, or reconfigurable logic devices that can reconfigure the connections inside the LSI or set up circuit sections inside the LSI can also be used for the same purpose.

All or part of the functions or operations of the cleaning route calculation system 1, cleaning unit 2, and terminal 3 shown in FIG. 1 can be executed by software processing. In this case, the software is recorded on one or more non-transitory storage media such as ROM, RAM, optical disk, or hard disk drive, and when the software is executed by a processor. Second, the software causes a processor and peripheral devices to perform specific functions within the software. Each of the cleaning route calculation system 1, the cleaning unit 2, and the terminal 3 includes one or more non-temporary recording media on which software is recorded, a processor, and necessary hardware devices. , for example, an interface. Software that implements the cleaning route calculation system 1 causes a computer to execute each step included in the flowchart shown in FIG. The software that implements the cleaning section 2 causes a computer to execute each step included in the flowchart shown in FIG. The software that implements the terminal 3 causes a computer to execute each step included in the flowchart shown in FIG.

The present disclosure can be used as a cleaning route determining device, for example, for efficient cleaning within a limited time at large-scale commercial facilities.

1 Cleaning route calculation system 2 Cleaning section 3 Terminal 10 Dust behavior calculation section 11 Drawing input section 12 Behavior calculation section 13 Dust map creation section 20 Attribute determination section 21 Attribute cluster calculation section 22 Attribute calculation section 30 Cleaning route calculation section 31 Dust cluster calculation Section 32 Cleaning route determination section 33, 54, 61 Communication section 40 Cleaning section 41 Dust collection section 42 Dust amount detection section 50 Drive section 51 Moving section 52 Control section 53 Self-position estimation section 60 Dust map display section 62 Display section 63 Input section

Claims (11)

a calculation condition input section that accepts calculation conditions including information inside the facility;
an analysis unit that analyzes the behavior of airflow and the behavior of particulates inside the facility based on the calculation conditions;
a map creation unit that creates a dust accumulation map showing one or more dust accumulation locations where dust accumulates and the amount of dust at the one or more dust accumulation locations within the facility, based on the results of the analysis; ,
a classification unit that selects one or more clusters from the one or more clusters inside the facility based on predetermined conditions;
a route calculation unit that determines a first route from a plurality of second routes;
an attribute cluster calculation unit that clusters a plurality of areas included inside the facility into one or more clusters by machine learning,
In each of the second plurality of routes, the cleaning unit passes through at least one dust accumulation place included in the selected one or more clusters among the one or more dust accumulation places within a predetermined time. is a route,
The selection by the classification unit is based on the characteristics of each of the one or more clusters,
If the number of entrances and exits of a room existing in the area of the selection target cluster is one, and the width of the passageway of the entrance and exit of the room is equal to or less than a second threshold, the classification unit classifies the selection target cluster. do not select,
Cleaning route determining device. The total amount of dust in all the dust accumulation locations included in the first path is the largest among the plurality of totals,
Each of the plurality of totals is a total of the amount of dust of all dust accumulation locations included in each of the second plurality of routes,
The cleaning route determining device according to claim 1. If the cluster to be selected is a store, the classification unit does not select the cluster to be selected;
The cleaning route determining device according to claim 1 or 2 . If the number of bends existing in the area of the selection target cluster is equal to or greater than a first threshold, the classification unit does not select the selection target cluster.
The cleaning route determining device according to claim 1 or 2 . If the minimum value of the width of the passage existing in the area of the selection target cluster is less than or equal to a third threshold value, the classification unit does not select the selection target cluster.
The cleaning route determining device according to claim 1 or 2 . If there is a step within the area of the selection target cluster, the classification unit does not select the selection target cluster;
The cleaning route determining device according to claim 1 or 2 . If the distance between the center of gravity of the area of the selection target cluster and the cleaning start position is equal to or greater than a fourth threshold, the classification unit does not select the selection target cluster;
The cleaning route determining device according to claim 1 or 2 . a calculation condition input section that accepts calculation conditions including information inside the facility;
an analysis unit that analyzes the behavior of airflow and the behavior of particulates inside the facility based on the calculation conditions;
a map creation unit that creates a dust accumulation map showing one or more dust accumulation locations where dust accumulates and the amount of dust at the one or more dust accumulation locations within the facility, based on the results of the analysis; ,
a classification unit that selects one or more clusters from the one or more clusters inside the facility based on predetermined conditions;
a route calculation unit that determines a first route from a plurality of second routes;
an attribute cluster calculation unit that clusters a plurality of areas included inside the facility into one or more clusters by machine learning,
In each of the second plurality of routes, the cleaning unit passes through at least one dust accumulation place included in the selected one or more clusters among the one or more dust accumulation places within a predetermined time. is a route,
The selection by the classification unit is based on the characteristics of each of the one or more clusters,
If the bias in the distribution of the fine particles within the area of the selection target cluster is equal to or greater than a fifth threshold, the classification unit does not select the selection target cluster;
Cleaning route determining device. If the frequency of people coming and going to the area of the selection target cluster is equal to or higher than a sixth threshold, the classification unit selects the selection target cluster.
The cleaning route determining device according to claim 1 or 2 . a calculation condition input step of accepting calculation conditions including information inside the facility;
an analysis step of analyzing the behavior of airflow and the behavior of particulates inside the facility based on the calculation conditions;
a map creation step of creating a dust accumulation map showing one or more dust accumulation locations where dust accumulates and the amount of dust at the one or more dust accumulation locations within the facility, based on the results of the analysis; ,
a classification step of selecting one or more clusters from one or more clusters inside the facility based on predetermined conditions;
a route calculation step of determining a first route from a plurality of second routes;
an attribute cluster calculation step of clustering a plurality of areas included inside the facility into one or more clusters by machine learning,
In each of the second plurality of routes, the cleaning unit passes through at least one dust accumulation place included in the selected one or more clusters among the one or more dust accumulation places within a predetermined time. is a route,
the selection by the classification step is based on characteristics of each of the one or more clusters;
If the number of entrances and exits of a room existing in the region of the selection target cluster is one, and the width of the passageway of the entrance and exit of the room is less than or equal to a second threshold value, the classification step do not select,
How to decide the cleaning route. a calculation condition input step of accepting calculation conditions including information inside the facility;
an analysis step of analyzing the behavior of airflow and the behavior of particulates inside the facility based on the calculation conditions;
a map creation step of creating a dust accumulation map showing one or more dust accumulation locations where dust accumulates and the amount of dust at the one or more dust accumulation locations within the facility, based on the results of the analysis; ,
a classification step of selecting one or more clusters from one or more clusters inside the facility based on predetermined conditions;
a route calculation step of determining a first route from a plurality of second routes;
an attribute cluster calculation step of clustering a plurality of areas included inside the facility into one or more clusters by machine learning,
In each of the second plurality of routes, the cleaning unit passes through at least one dust accumulation place included in the selected one or more clusters among the one or more dust accumulation places within a predetermined time. route,
the selection by the classification step is based on characteristics of each of the one or more clusters;
If the bias in the distribution of the fine particles within the area of the selection target cluster is equal to or greater than a fifth threshold, the classification step does not select the selection target cluster;
How to decide the cleaning route.

Images